Forum : ARM
Original Post
Post Information | Post |
---|---|
October 14, 2010 - 3:38pm
|
Leaving Ride running for more than (5-6 hours) while debugging my target (STM32F) causes it to be unresponsive. Task Manager shows that the amount of memory usage has ballooned to approx. four times what it was when debugging was first started. This only becomes a problem when trying to debug some intermittent issues that would benefit from Ride's debugging capabilities. I'm using Ride 7.30.10.0159 with the 7.30.10.0169 patch and RKit-ARM 1.26.10.0130. |
Hi,
We would like to help you as quickly as possible on this issue, but we need to have more information about the problem before we can help further.
In order to understand the issue you are facing, we need the following information about your configuration:
- Version of Windows: At a command prompt type the following command: "systeminfo.exe", which gives you a system report. We need the full "Operating System" and "Version" from the report.
- Which processor derivative are you debugging, and on which hardware platform (can you send us the schematics)?
Concerning your particular problem:
- Provide the steps you followed to raise the error.
- Does the problem occur when using the examples given with Ride7 (from the Ride7 installation directory, default being "C:\Program Files\Raisonance\Ride\Examples\ARM") and a Raisonance hardware (such as a REva board)?
Regards,
Hi,
We reproduced your problem and we are working on fixing it.
If you are working with STM8, you can workaround the problem by disabling Read-On-The-Fly in the debug options.
If you are working on ARM there is no workaround for now. We will propose a patch for disabling ROTF if the fixing of the root problem appears to take time.
Best Regards,
Vincent
Unfortunately, I am using ARM - STM32F103VET6 to be exact.
What does ROTF give you in debug?
Let me know when something becomes available.
Thanks.
Hi,
ROTF is the debugger updating the content of memory views, peripheral views and watches while the target CPU is executing the debugged application. (running) It is available on new devices like STM8 and Cortex, but not on old ones like ST7, ARM7, etc.
Anyways, here is the test-version of a patch that should solve the problem (no need to disable ROTF):
ftp://www.raisonance.com/temp/Ride7/patch_101019_memleak_rotf.zip
Please let us know if it works for you.
We need some customer feedback before releasing it officially.
Best Regards,
Vincent
My coworkers and I have been using the patch for a couple days and it seems to work. The memory footprint over extended use is definitely improved.
I do have a couple notes -
- I notice that operations like opening and closing register and peripheral views seem to permanently increase the memory footprint (I watched it stay at the same level for approx. 5 min. anyway). The increase is very small, but why is the increase sustained?
- Minimizing and restoring the Ride windows after launching debug seems to greatly decrease the memory footprint. I understand why minimizing decreases the footprint, but why is it dramatically less (approx. 80%) when I restore it to approx. the same state?
Hi,
Thank you for your feedback.
We will inquire on your new reports. But they are not critical, (they don't prevent you fron using the tools) so the fixes will not be available as a patch. They will included in a future official release.
Best Regards,
Vincent